A couple of interesting things about the journey from the press release to the story: the press release says "associated with", but the summary says "impacts on", and the headline says "critical for preventing"! Also, they pick "screen time and sleep" from among five factors, the others were less tobacco smoking, more physical exercise and more (yes) alcohol.

I don't have a problem with causal claims from these kinds of methods in principle, if the regression is well done with good controls. But I'm not at all convinced that they can isolate self-reported sedentary screen time reliably from self-reported physical activity.

https://bmcmedicine.biomedcentral.com/track/pdf/10.1186/s129...